Emergency Responder Stationing

Published:

πŸ“Œ Key Contributions

  • We introduce a novel solution approach using Deep Reinforcement Learning and Combinatorial Optimization techniques to enable real-time decision-making.
  • We use DDPG to train agents for performing redistribution actions (city-scale) and reallocation actions (region-scale).
  • We utilize a Transformer-based actor to handle variable numbers of responders and depots during region-level reallocation.
  • We map continuous actions exactly to discrete actions using combinatorial optimization (min-cost flow + max-weight matching), preserving gradient flow while ensuring feasibility.
  • We signal the performance of high-level actions through low-level critics.
  • Our trained DRL agents achieve 1000x faster decision-making while reducing response times to between 5 and 13 seconds on real-world datasets.

πŸ” High-Level Overview of the SOTA Approach with Hierarchical Coordination

High-Level Process

This diagram illustrates our state-of-the-art hierarchical coordination framework that combines queuing based city-scale redistributions and MCTS based region-level reallocations of responders.


🧠 Region-Level Reallocation via DDPG Training

Region-Level Training

We leverage DDPG to train agents that perform region-level reallocation of responders, enabling efficient adaptation to changing demand at a broader geographic scale.


πŸ™οΈ City-Level Redistribution via DDPG Training

City-Level Training

At the city scale, DDPG is used to train agents for fine-grained redistribution of responders, allowing precise real-time response in dense urban environments.


πŸ“ Publication

Published as a full paper at ICML 2024 β€” β€œMulti-Agent Reinforcement Learning with Hierarchical Coordination for Emergency Responder Stationing.” [OpenReview]


πŸ’» Code & Data

Reproducible code, training scripts, and Nashville & Seattle datasets: [Code & Data]


πŸŽ₯ 3-Minute Overview

Summarising the challenges, solution approach, and results: [Short Video]